Roman concrete survives in ancient harbours and monumental buildings, often under demanding conditions. Parts of ancient harbour structures have resisted seawater for nearly two thousand years, while the Pantheon’s enormous dome still stands over Rome, as it has since Hadrian’s reign. Their durability has several explanations because Roman builders used different mixtures for different purposes. A harbour wall submerged in salt water and a sheltered dome have had very different histories of use, exposure and repair. The evidence does not point to a single Roman recipe.
The mortar binding the rubble is one part of the explanation. Roman builders often mixed lime with volcanic ash or other reactive minerals. Such mixtures could harden in wet conditions because their chemistry differed from ordinary lime mortar. Some Roman concrete also contains small, calcium-rich fragments left within the binder. A 2023 study examined whether these fragments could help close cracks, while evidence from an unfinished construction site at Pompeii later showed how certain Roman workers actually prepared their mortar.
Roman builders selected aggregates, moved building materials and placed wet mixtures inside temporary wooden forms. They also adjusted the mass of a structure to suit the forces it had to bear. Modern chemistry helps explain reactions whose effects the Romans could observe without understanding their molecular causes. Modern engineers are investigating these reactions for practical use. Roman concrete could still deteriorate or fail, and adopting its methods would not eliminate the emissions from cement production.
What the Romans meant by concrete
The Latin term opus caementicium describes construction in which mortar holds fragments of stone or other rubble together. The caementa were those fragments, the ingredients that formed the bulk of the mass, rather than anything equivalent to a bag of modern cement. Roman walls might have a rubble core faced with brick or dressed stone, giving the structure both shape and a finished surface. Builders also used variations of the material in vaults, foundations, baths, cisterns and harbour works. The label ‘Roman concrete’ covers mixtures with varying aggregates, binder ratios, raw materials and setting conditions.
Lime mortar begins with limestone, which is largely calcium carbonate. Heating it releases carbon dioxide and produces calcium oxide, or quicklime. Adding water produces calcium hydroxide, known as slaked lime, in a reaction that also gives off heat. Ordinary lime mortar hardens as that hydroxide absorbs carbon dioxide from the air and returns to calcium carbonate. Deep inside masonry, or under water, the process can be very slow. By combining lime with appropriate volcanic ash or similar reactive materials, Roman builders made hydraulic mortar. Its hardening did not depend entirely on exposure to air, so it could work in damp places and below the surface of the sea.
The ash reacted chemically with the lime. Suitable volcanic material contains reactive silica and alumina, often within volcanic glass. In the alkaline conditions created by lime, these ingredients can dissolve and form calcium-silicate-hydrate and related calcium-aluminosilicate compounds. The products bind aggregates and can make the passages through which water enters the mortar smaller. Such a reactive material is called a pozzolan, a word derived from Pozzuoli near Naples, whose volcanic deposits became famous in antiquity. Not all volcanic rocks react equally well, and Roman builders sometimes used other locally available substances.
In the late first century BCE, the Roman architect
Harbour construction made hydraulic mortar especially useful. A coast without a naturally sheltered inlet needed piers and breakwaters to protect ships from waves and provide places to land. Workers could drive timber piles into the seabed, brace a wooden enclosure and fill it with a mixture of stone, lime and volcanic ash. As the mortar hardened, the enclosed material formed a solid body in the shape of the temporary framework. Building in rough water remained difficult, but concrete allowed substantial underwater works without having to assemble every section from closely fitted, dressed masonry blocks. Vitruvius describes the operation in On Architecture 5.12. For sea works, he recommends volcanic powder from between Cumae and the promontory of Minerva, mixed at a ratio of two parts powder to one part lime. He goes on to explain how workers should drive and brace stakes, prepare the foundation and fill the space within the timber enclosure. Elsewhere he discusses alternatives where the right powder or workable conditions were lacking. The instructions show why the binder mattered to the method: builders could set a large mass in water rather than trying to keep the whole construction site dry. Suitable volcanic ash was sometimes transported across the Mediterranean. The Roman Maritime Concrete Study, or ROMACONS examined surviving harbour concrete through survey, archaeological sampling and petrographic study. The researchers found Italian pozzolanic material at Caesarea Maritima, the great harbour built for Herod on the eastern Mediterranean coast. They also studied Chersonisos on Crete, where preliminary results suggested an Italian volcanic component rather than ash from Santorini. Other origins remained possible, and the researchers acknowledged the limits of their samples. The samples document long-distance movement of specialist building materials. They do not establish a centrally managed supply network for every Roman harbour. Builders chose material according to local geology, expense, labour and what a structure needed to withstand. Campanian ash might justify a long voyage for an ambitious maritime project, while an inland site could use more accessible stone, ash or even crushed ceramic material. Mortars were not identical across the empire, and different phases of a single harbour might reflect rebuilding, repair or changing choices of aggregate and binder.Volcanic ash and the engineering of Roman harbours
How seawater could help strengthen a Roman pier
Salt water attacks much modern reinforced concrete when chlorides reach the steel bars, causing corrosion that expands and breaks surrounding material. Roman marine concrete normally had no comparable embedded steel, so it avoided that particular mechanism. Seawater could also behave differently within its lime-and-volcanic-ash binder. Flowing through pores and cracks, it encountered reactive minerals and sometimes contributed to the formation of new ones instead of simply destroying the masonry.
Marie Jackson and her colleagues studied ancient sea concrete using microscopy and synchrotron X-ray methods. Their 2017 study in American Mineralogist identified phillipsite, a zeolite, and aluminous tobermorite within some Roman harbour samples. These minerals occurred in spaces in the binder and in altered volcanic grains, forming as seawater and alkaline fluids interacted with volcanic material over long periods at relatively low temperatures. Their growth can help bind parts of the concrete and impede the spread of fractures. The composition of other piers has to be established from their own samples.
These reactions occurred over different periods of the concrete’s history. Lime reacted with the pozzolanic material soon after construction to create the initial hydraulic binder. Over centuries, water entering the pores could dissolve further components of volcanic glass and allow additional minerals to crystallise. Jackson’s team reconstructed aspects of that process from crystals inside tiny cavities and altered particles. Some of the concrete continued to change chemically long after it first hardened, with reactions that in certain conditions could improve its durability.
In Natural History 35.166, Pliny the Elder praised the powder of the Puteoli district for becoming solid in water and resisting the sea. His description records the reputation of a valuable building material. It cannot establish that every Roman harbour structure increased in strength with age. Some structures failed, and much of the vanished concrete is unavailable for chemical comparison. The works that survive are only part of the ancient building record.
The small white fragments that changed the explanation
Recent research has also turned to Roman concrete that never met seawater. In 2023, Linda Seymour, Janille Maragh, Paolo Sabatini, Michel Di Tommaso, James Weaver and Admir Masic published a study of lime-rich fragments in mortar from the city wall of Privernum near Rome. The mortar contained small white calcium-rich pieces, often called lime clasts, scattered through a darker matrix. Earlier explanations had sometimes treated these fragments as evidence of badly mixed material or lime that had not been properly prepared. The researchers considered whether the fragments resulted from a deliberate way of mixing and could serve a useful purpose after construction.
They examined the inclusions and surrounding mortar with scanning electron microscopy, chemical mapping, X-ray diffraction and Raman spectroscopy. The evidence supported a process in which builders introduced quicklime directly, either alongside or instead of lime that had already been slaked with water. Quicklime reacts strongly when wetted and releases heat, so hot mixing changes the conditions under which the mortar forms. It can leave distinctive calcium-rich pieces within the hardened material. The team proposed that these might remain a store of reactive calcium capable of influencing later cracks.
The proposed crack-sealing process has several steps. Temperature changes, movement and other stresses can create narrow fractures. Water reaching a lime-rich fragment may dissolve calcium compounds, which can then react with dissolved carbon dioxide to precipitate calcium carbonate inside the opening. Reactions involving nearby pozzolanic grains might add further material. Whether a crack closes depends on its size, how water passes through it and whether the ingredients required for these reactions remain available. A structure with severe damage cannot be expected to repair itself simply because it contains lime.
The 2023 researchers tested this possibility in the laboratory. They made modern mortars based on their understanding of the ancient mixture, deliberately fractured them and sent water through the cracks. Under their test conditions, openings approximately half a millimetre wide or smaller sometimes sealed within weeks and no longer allowed water through. Control samples without the relevant quicklime inclusions did not respond in the same way. The results support a mechanism that may be useful to engineers. They cannot stand in for two thousand years of controlled observation, or prove that the same mechanism preserved every aqueduct, harbour wall and the Pantheon.
Pompeii shows how builders actually mixed the ingredients
The Privernum study also prompted researchers to look for direct evidence of how the mortar was prepared. Ancient descriptions often seem to recommend slaking lime before incorporating it into mortar. Finished masonry can suggest how mortar formed, but an abandoned building site might preserve the ingredients and working stages needed to identify the procedure.
Pompeii supplied such a site because the eruption of Vesuvius in 79 CE interrupted building work in Regio IX. Excavations exposed unfinished masonry and heaps of raw materials still awaiting use. In a study published on 9 December 2025 in Nature Communications, Ellie Vaserman and colleagues compared the material piles with completed walls and those in the middle of construction. Their chemical and microscopic findings indicate that quicklime and pozzolanic material were mixed together dry before water was introduced.
The workers had left evidence for several stages of the job. Researchers could examine prepared ingredients and mortar already incorporated into walls, rather than inferring the entire process from white particles in finished concrete. They identified calcium-rich pieces, altered volcanic grains and mineral changes associated with hot mixing. The material abandoned in 79 CE thus linked a method of preparation to the traces it left inside the mortar, something a standing monument alone cannot reveal.
Pompeii provides strong evidence that Roman builders sometimes used hot mixing, without proving that everyone did so. Other teams may have slaked quicklime first, added it during mixing or used both practices according to their ingredients and building needs. Vitruvius described the methods he recommended to readers; the excavated site captures decisions made by workers on a particular project. Read together, they show why neither a written prescription nor a single building site establishes a universal Roman method.
The Pompeii evidence also clarifies what researchers mean by “self-healing.” Researchers observed calcium moving from lime-rich material into adjacent areas and new carbonate minerals forming. Similar reactions may close tiny pores and cracks under the right conditions. That is a microscopic form of repair, not evidence that an ancient wall could reconstruct itself after major structural damage. Roman concrete remained vulnerable to erosion, excessive loads and other causes of failure despite its capacity for unusually long survival.
Why the Pantheon’s dome is still standing
The Pantheon presents a very different engineering problem from a pier. Its concrete dome, completed during the reign of Hadrian in the second century CE, extends roughly 43 metres across without a modern cage of reinforcing steel. A huge circular drum supports it, aided by substantial foundations, thick walls and relieving arches that distribute the loads. The building has remained in use and received protection and repairs over centuries. Its survival reflects that continuing history as well as the skill with which it was constructed.
Roman builders made the dome progressively lighter as it rose. The base is thick and includes relatively heavy aggregates; higher up, the mixture contains lighter volcanic stone, and the thickness decreases toward the central opening, the oculus. Recessed coffers also shape the interior. A dome presses down on its supports and creates outward forces, so reducing weight near the top and providing sufficient support below were essential decisions. The structure’s geometry, building sequence and material choices belong to the same explanation as its mortar.
We cannot assume that hot mixing, as reconstructed from mortar at Privernum, accounts for the Pantheon’s endurance. The 2023 study did not take samples from the dome and did not determine whether lime clasts occur there in the relevant form or quantities. Land-based Roman concrete could contain lime, ash and locally chosen aggregates, but a dome sheltered from seawater behaves differently from a submerged pier. The dome’s design, its materials and centuries of maintenance all matter to its survival.
Nor does the Pantheon establish that contemporary builders could not construct a similarly large dome without steel. They generally work under different building codes and design requirements, with reinforced systems chosen to control cracking and bear tensile stresses. A Roman mass-concrete technique might suit particular projects, but replacing the concrete in a modern bridge deck or tall building with an ancient mixture would ignore the forces, safety standards and expected performance of those structures.
The climate question: can Roman ideas reduce cement emissions?
Modern cement creates a different set of concerns. Most concrete today contains cement made with Portland clinker, produced by heating limestone and other minerals to about 1,450°C. The process releases carbon dioxide from the fuel burned in the kiln and from the limestone itself as calcium carbonate decomposes. Cleaner heat would address only the first source unless process emissions were also captured or the chemistry altered. Cement production is therefore an important industrial contributor to greenhouse gases, though its stated percentage of global emissions depends on which year and accounting method are used.
Roman builders did not produce Portland clinker. Some mixed their binders with large proportions of volcanic ash, an approach that points toward replacing part of today’s high-emission clinker with suitable supplementary materials. Ancient lime production was not free of emissions: limestone still had to be heated to make quicklime, releasing carbon dioxide and consuming fuel. Shipping heavy aggregates and building thick masonry also required resources. Meaningful environmental comparisons must cover the construction and useful life of a whole structure, rather than setting two kiln temperatures side by side.
One modern approach is limestone-calcined clay cement (LC3). Research at EPFL on limestone-calcined clay cement, often called LC3 describes blends combining a reduced share of clinker with ground limestone and clay heated at considerably lower temperatures than those used to make Portland clinker. Some blends can meet the requirements of ordinary structural construction. EPFL has described potential reductions of around 40 percent in emissions compared with conventional cement for particular formulations. Actual savings depend on the clay and other raw materials, the energy and transport involved, the mix and the strength and durability demanded of the finished concrete.
A 2025 EPFL-linked study of low-binder LC3 structural concretes considered another way to cut emissions: use less binder where the structure can meet its performance requirements without it. Prescriptive minimum-cement rules sometimes lead builders to add more than is technically needed. Combining efficient structural design with reduced clinker content can address both problems. Researchers are not trying to recreate a Roman pier in every modern building. The trials seek adequate strength, workability and service life with less carbon-intensive material. Roman pozzolanic chemistry provides one line of inquiry.
What engineers can borrow, and what they still need to test
In some environments, the hardened binder continues to react with its surroundings. Samples of Roman harbour concrete show how mineral reactions in the presence of seawater can proceed over long periods. Research on hot mixing suggests that stored calcium in some inclusions may help close fine cracks. Modern scientists hope to use these findings to make binders less permeable, slow the movement of damaging substances and extend the time between repairs. Such improvements would be valuable in coastal buildings, foundations and other places where moisture steadily challenges the structure.
Not all enduring Roman hydraulic works relied on concrete. The Pont du Gard in southern France survives principally as precisely cut and assembled stone masonry, so it would be misleading to cite it as proof of the chemistry of ancient marine concrete. The account by Sextus Julius Frontinus of the management of Rome’s aqueducts also shows the role of inspection, repair and administration. A water system lasted because people maintained it, as well as because its builders chose effective materials and methods.
Volcanic ash is a broad category, and two deposits may differ sharply in how they react. Modern engineers must test natural pozzolans, calcined clays and other supplementary materials for particle size, reactive silica and alumina, compatibility with the binder and performance in local conditions. Working directly with quicklime requires control of the heat it produces when wetted. A mixture that closes a narrow crack in a laboratory may still be unsuitable for a building if it sets too slowly, shrinks excessively or lacks the required workability or strength.
In August 2026, MIT described the work of Dmat. Dmat is a company co-founded by Admir Masic to develop Roman-inspired concrete technology. Its stated emissions and performance figures refer to particular products and tests. They do not amount to independent evidence that Roman methods have resolved the concrete industry’s environmental and durability problems. New binders still require comparable long-term field results, independent assessment and compliance with the standards governing structures where failure could be dangerous.
The ancient evidence is also a selective record of structures that survived. A harbour pier that survived may have benefited from unusually favourable materials, exposure, maintenance or circumstances, while unsuccessful neighbours have disappeared. Researchers cannot sample a demolished mass of concrete with the same detail as an intact monument. Results from a few Roman sites cannot represent every century, province or type of structure. This is why Pompeii’s abandoned worksite matters: unlike a surviving wall, it preserves evidence for how the material was prepared before anyone knew how long it would last.
Roman concrete lasted where engineering choices suited the job and, in some cases, where its minerals continued reacting in ways that helped preserve it. Ancient builders knew how their materials behaved in practice, even though they lacked today’s chemical explanations. Their choice of ingredients for particular conditions remains instructive. Modern engineers can extend that approach by measuring how a binder performs in its intended environment over years and decades, instead of expecting the age of the Pantheon to certify a new concrete mixture.
Discussion
Comments